Biology Reference
In-Depth Information
insufficient. To repeat, the entire skeleton should be inventoried for any and all pathological
changes and, if feasible, include radiographic or CT (computed tomography) imaging and
molecular data (e.g., DNA identification of the pathogen). Comprehensive paleopathological
inventorying certainly demands an objective, non-biased scoring and quantification protocol.
Scoring criteria are available but note that there are still unresolvedproblems of reproducibility
(e.g.,
Waldron and Rogers, 1991; Jacobi and Danforth, 2002; Grauer, 2008; Stodder, 2012
).
Quantification
If the reader is interested in undertaking a paleopathological study, understanding how to
quantify the data will be essential as the strategy here differs from that of other areas in skel-
etal biology. As a result, the most difficult aspect of diagnostics is probably quantification.
Pathologies are not necessarily a simple presence
e
absence phenomenon. There will likely
be a continuous grade from normal to abnormal and often it is difficult to determine the
threshold of abnormal. This is certainly true of EC (entheseal change). Even if a threshold
for abnormal is determined, severity will also be manifested on a progressive continuum.
It is admittedly difficult to force a continuous phenomenon into discrete categories. Addi-
tionally, severity may vary between populations and may fundamentally affect the quantifi-
cation of disease prevalence. For example, an observer's characterization of severity as mild
in one sample may be characterized by another observer in a second sample as being at
a threshold-normal level. This difference in scoring criteria is called observer error. It may
reflect the differences between multiple observers (interobserver error) or within a single
observer over a large skeletal sample (criteria drift).
Observers must particularly be vigilant against criteria drift occurring throughout the
process of data collection as one gradually (almost subconsciously) changes their decision-
making process with application of the scoring protocol as the number of skeletons (and
therefore variety of features/pathologies) they have seen increases. One way to minimize
observer error is to do a blind study in which two observers independently score a given
pathology and compare the results. Even if scoring is equivalent, it should be spot-checked
periodically for criteria drift. Single observers should rescore the first twenty or so cases at
a later date to test for criteria drift. If changes in the threshold of presence or severity are
detected, the sample should be spot-checked until no discrepancy occurs.
Scoring Protocol
There are several standards and protocols available complete with exemplary images.
These include the aforementioned resources in the earlier section on reference materials.
There are also software programs such as Osteoware,
9
a data-entry program that was devel-
oped by the Smithsonian Institution that helpfully includes an image module to facilitate
scoring (
http://osteoware.si.edu/guide/pathology-module
)
. However, this does not mean
that ambiguities will not occur (e.g.,
Jacobi and Danforth, 2002
). Therefore, before under-
taking any given project, it is always a good idea to survey the literature for precise and
reproducible scoring criteria. Publications and reports will contain a Materials and Methods
9
Downloadable free standardized skeletal documentation software from the Smithsonian Institution:
